Method
Setting
Code
Fl-bg
Fl-fg
Fl-all
Density
Runtime
Environment
1
ScaleRAFTRBO
code
2.27 %
5.63 %
2.83 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (C/C++)
2
SplatFlow3D
code
2.27 %
6.02 %
2.89 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
3
GAOSF
2.08 %
7.37 %
2.96 %
100.00 %
1 s
GPU @ 2.5 Ghz (Python + C/C++)
4
CamLiRAFT
code
2.08 %
7.37 %
2.96 %
100.00 %
1 s
GPU @ 2.5 Ghz (Python + C/C++)
H. Liu, T. Lu, Y. Xu, J. Liu and L. Wang: Learning Optical Flow and Scene Flow
with Bidirectional Camera-LiDAR Fusion . TPAMI 2023.
5
OAMaskFlow
2.10 %
7.80 %
3.05 %
100.00 %
0.5 s
GPU @ 2.5 Ghz (Python)
6
CamLiFlow
code
2.31 %
7.04 %
3.10 %
100.00 %
1.2 s
GPU @ 2.5 Ghz (Python + C/C++)
H. Liu, T. Lu, Y. Xu, J. Liu, W. Li and L. Chen: CamLiFlow: Bidirectional Camera-LiDAR Fusion for Joint
Optical Flow and Scene Flow Estimation . CVPR 2022.
7
RAFT-3D++
2.09 %
8.55 %
3.16 %
100.00 %
0.5 s
1 core @ 2.5 Ghz (Python)
8
DDVM
2.90 %
5.05 %
3.26 %
100.00 %
S. Saxena, C. Herrmann, J. Hur, A. Kar, M. Norouzi, D. Sun and D. Fleet: The Surprising Effectiveness of Diffusion Models for Optical Flow and Monocular Depth Estimation . NeurIPS 2023.
9
CamLiRAFT-NR
code
2.76 %
6.78 %
3.43 %
100.00 %
1 s
GPU @ 2.5 Ghz (Python + C/C++)
H. Liu, T. Lu, Y. Xu, J. Liu and L. Wang: Learning Optical Flow and Scene Flow
with Bidirectional Camera-LiDAR Fusion . arXiv preprint arXiv:2303.12017 2023.
10
M-FUSE
code
2.66 %
7.47 %
3.46 %
100.00 %
1.3 s
GPU
L. Mehl, A. Jahedi, J. Schmalfuss and A. Bruhn: M-FUSE: Multi-frame Fusion for Scene Flow Estimation . Proc. Winter Conference on Applications of Computer Vision (WACV) 2023.
11
RigidMask+ISF
code
2.63 %
7.85 %
3.50 %
100.00 %
3.3 s
GPU @ 2.5 Ghz (Python)
G. Yang and D. Ramanan: Learning to Segment Rigid Motions from Two
Frames . CVPR 2021.
12
CroCo-Flow
code
3.18 %
5.94 %
3.64 %
100.00 %
3s
NVIDIA A100
P. Weinzaepfel, T. Lucas, V. Leroy, Y. Cabon, V. Arora, R. Br\'egier, G. Csurka, L. Antsfeld, B. Chidlovskii and J. Revaud: CroCo v2: Improved Cross-view Completion
Pre-training for Stereo Matching and Optical Flow . ICCV 2023.
13
SoloFlow
3.34 %
5.35 %
3.68 %
100.00 %
1 s
1 core @ 2.5 Ghz (C/C++)
14
SOLOFlow
3.47 %
5.41 %
3.79 %
100.00 %
1 s
1 core @ 2.5 Ghz (C/C++)
15
CCMR+
code
3.39 %
6.21 %
3.86 %
100.00 %
1.5 s
GPU @ 2.5 Ghz (Python)
A. Jahedi, M. Luz, M. Rivinius and A. Bruhn: CCMR: High Resolution Optical Flow Estimation via Coarse-to-Fine Context-Guided Motion Reasoning . WACV 2024.
16
MemFlow-T
code
3.44 %
6.09 %
3.88 %
100.00 %
Q. Dong and Y. Fu: MemFlow: Optical Flow Estimation and
Prediction with Memory . Proceedings of the IEEE/CVF Conference
on Computer Vision and Pattern Recognition 2024.
17
RAFT-it+_RVC
code
3.62 %
5.33 %
3.90 %
100.00 %
0.14 s
1 core @ 2.5 Ghz (Python)
D. Sun, C. Herrmann, F. Reda, M. Rubinstein, D. Fleet and W. Freeman: Disentangling Architecture and Training for
Optical Flow . ECCV 2022.
18
RAFT-OCTC
3.72 %
5.39 %
4.00 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
J. Jeong, J. Lin, F. Porikli and N. Kwak: Imposing Consistency for Optical Flow
Estimation (Qualcomm AI Research) . CVPR 2022.
19
mf+
3.80 %
5.37 %
4.07 %
100.00 %
1 s
1 core @ 2.5 Ghz (C/C++)
20
MotionFlow
3.95 %
4.74 %
4.08 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (C/C++)
21
MemFlow
code
3.67 %
6.27 %
4.10 %
100.00 %
Q. Dong and Y. Fu: MemFlow: Optical Flow Estimation and
Prediction with Memory . Proceedings of the IEEE/CVF Conference
on Computer Vision and Pattern Recognition 2024.
22
SF2SE3
code
3.17 %
8.79 %
4.11 %
100.00 %
2.7 s
GPU @ >3.5 Ghz (Python)
L. Sommer, P. Schröppel and T. Brox: SF2SE3: Clustering Scene Flow into SE (3)-Motions via Proposal and Selection . DAGM German Conference on Pattern Recognition 2022.
23
MS_RAFT+_corr_RVC
code
3.83 %
5.71 %
4.15 %
100.00 %
0.65 s
GPU @ 2.5 Ghz (Python + C/C++)
A. Jahedi, M. Luz, M. Rivinius, L. Mehl and A. Bruhn: High Resolution Multi-Scale RAFT . International Journal of Computer Vision (IJCV) 2023.
24
app+mo1
3.95 %
5.21 %
4.16 %
100.00 %
1 s
1 core @ 2.5 Ghz (C/C++)
25
FlowDiffuser
code
3.68 %
6.64 %
4.17 %
100.00 %
0.4 s
1 core @ 2.5 Ghz (Python)
A. Luo, X. Li, F. Yang, J. Liu, H. Fan and S. Liu: FlowDiffuser: Advancing Optical Flow
Estimation with Diffusion Models . In Proc. CVPR 2024.
26
DIP
code
3.86 %
5.96 %
4.21 %
100.00 %
0.15 s
1 core @ 2.5 Ghz (Python)
Z. Zheng, N. Nie, Z. Ling, P. Xiong, J. Liu, H. Wang and J. Li: DIP: Deep Inverse Patchmatch for High-
Resolution Optical Flow . Proceedings of the IEEE/CVF
Conference on Computer Vision and Pattern
Recognition 2022.
27
StreamFlow
3.91 %
5.91 %
4.24 %
100.00 %
0.08 s
1 core @ 2.5 Ghz (Python)
28
ACR-Net
4.06 %
5.16 %
4.25 %
100.00 %
0.08 s
1 core @ 2.5 Ghz (C/C++)
29
ScaleFlow++
3.95 %
5.76 %
4.25 %
100.00 %
0.1 s
GPU @ 2.5 Ghz (Python)
30
MonoFusion
3.93 %
5.97 %
4.27 %
100.00 %
0.7 s
GPU @ 2.5 Ghz (Python)
31
RAFT-3D
3.39 %
8.79 %
4.29 %
100.00 %
2 s
GPU @ 2.5 Ghz (Python + C/C++)
Z. Teed and J. Deng: RAFT-3D: Scene Flow using Rigid-Motion
Embeddings . arXiv preprint arXiv:2012.00726 2020.
32
CCAFlow
4.05 %
5.56 %
4.30 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
33
SEA-RAFT(L)
4.08 %
5.37 %
4.30 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (Python)
34
Promotion
4.02 %
5.73 %
4.30 %
100.00 %
0.25 s
1 core @ 2.5 Ghz (Python)
35
sdflowex
3.97 %
5.96 %
4.30 %
100.00 %
1.2 s
1 core @ 2.5 Ghz (C/C++)
36
ap+m
3.96 %
6.11 %
4.32 %
100.00 %
1 s
1 core @ 2.5 Ghz (C/C++)
37
SGFlow
4.04 %
5.82 %
4.34 %
100.00 %
0.15 s
1 core @ 2.5 Ghz (Python)
38
ProtoFormer
4.05 %
5.81 %
4.34 %
100.00 %
0.2 s
1 core @ 2.5 Ghz (Python)
39
DF-Flow
4.10 %
5.60 %
4.35 %
100.00 %
0.2 s
1 core @ 2.5 Ghz (Python)
40
GMFlow_RVC
code
4.16 %
5.67 %
4.41 %
100.00 %
0.2 s
GPU (Python)
H. Xu, J. Zhang, J. Cai, H. Rezatofighi, F. Yu, D. Tao and A. Geiger: Unifying Flow, Stereo and Depth
Estimation . arXiv preprint arXiv:2211.05783 2022.
41
AnyFlow
4.15 %
5.76 %
4.41 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (Python)
H. Jung, Z. Hui, L. Luo, H. Yang, F. Liu, S. Yoo, R. Ranjan and D. Demandolx: AnyFlow: Arbitrary Scale Optical Flow with
Implicit Neural Representation . arXiv preprint arXiv:2303.16493 2023.
42
ASFlow
4.17 %
5.68 %
4.42 %
100.00 %
0.4 s
1 core @ 2.5 Ghz (C/C++)
43
FP-TTC
4.23 %
5.71 %
4.48 %
100.00 %
0.15 s
1 core @ 2.5 Ghz (C/C++)
44
GMFlow+
code
4.27 %
5.60 %
4.49 %
100.00 %
0.2 s
GPU (Python)
H. Xu, J. Zhang, J. Cai, H. Rezatofighi, F. Yu, D. Tao and A. Geiger: Unifying Flow, Stereo and Depth
Estimation . arXiv preprint arXiv:2211.05783 2022.
45
ScaleRAFT
code
4.45 %
4.76 %
4.50 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (C/C++)
46
SeparableFlow
code
4.25 %
5.92 %
4.53 %
100.00 %
0.5 s
GPU
F. Zhang, O. Woodford, V. Prisacariu and P. Torr: Separable Flow: Learning Motion Cost
Volumes for Optical Flow Estimation . Proceedings of the IEEE/CVF
International Conference on Computer Vision 2021.
47
HD-Flow
4.10 %
6.67 %
4.53 %
100.00 %
0.2 s
1 core @ 2.5 Ghz (Python)
48
f
4.38 %
5.61 %
4.59 %
100.00 %
1 s
1 core @ 2.5 Ghz (C/C++)
49
DFFlow
4.30 %
6.08 %
4.60 %
100.00 %
1 s
1 core @ 2.5 Ghz (Python)
50
KPA-Flow
4.17 %
6.77 %
4.60 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
A. Luo, F. Yang, X. Li and S. Liu: Learning Optical Flow With Kernel Patch
Attention . In Proc. CVPR 2022.
51
SplatFlow
code
4.26 %
6.34 %
4.61 %
100.00 %
0.1 s
GPU @ 2.5 Ghz (Python)
B. Wang, Y. Zhang, J. Li, Y. Yu, Z. Sun, L. Liu and D. Hu: SplatFlow: Learning Multi-frame Optical Flow via Splatting . International Journal of Computer Vision 2024.
52
MatchFlow(G)
code
4.33 %
6.11 %
4.63 %
100.00 %
0.3 s
GPU (Python)
Q. Dong, C. Cao and Y. Fu: Rethinking Optical Flow from Geometric Matching Consistent
Perspective . Proceedings of the IEEE/CVF Conference on Computer
Vision and Pattern Recognition 2023.
53
RPKNet
code
4.63 %
4.69 %
4.64 %
100.00 %
0.6 s
GPU @ 2.5 Ghz (Python)
H. Morimitsu, X. Zhu, X. Ji and X. Yin: Recurrent Partial Kernel Network for
Efficient Optical Flow Estimation . AAAI 2024.
54
TrepFlow
4.42 %
5.91 %
4.66 %
100.00 %
0.04 s
GPU @ 2.5 Ghz (Python)
55
FlowFormer
code
4.37 %
6.18 %
4.68 %
100.00 %
0.3 s
GPU (Python)
Z. Huang, X. Shi, C. Zhang, Q. Wang, K. Cheung, H. Qin, J. Dai and H. Li: FlowFormer: A Transformer Architecture for
Optical Flow . European conference on computer
vision 2022.
56
SSTM_T [MV]
4.39 %
6.40 %
4.72 %
100.00 %
0.4 s
GPU @ 2.5 Ghz (C/C++)
F. Ferede and M. Balasubramanian: SSTM: Spatiotemporal recurrent
transformers for multi-frame optical flow
estimation . Neurocomputing 2023.
57
MatchFlow(R)
code
4.51 %
5.78 %
4.72 %
100.00 %
0.26 s
GPU (Python)
Q. Dong, C. Cao and Y. Fu: Rethinking Optical Flow from Geometric
Matching Consistent
Perspective . Proceedings of the IEEE/CVF
Conference on Computer
Vision and Pattern Recognition 2023.
58
UberATG-DRISF
3.59 %
10.40 %
4.73 %
100.00 %
0.75 s
CPU+GPU @ 2.5 Ghz (Python)
W. Ma, S. Wang, R. Hu, Y. Xiong and R. Urtasun: Deep Rigid Instance Scene Flow . CVPR 2019.
59
AGFlow
code
4.16 %
7.59 %
4.73 %
100.00 %
0.2 s
8 cores @ 2.5 Ghz (Python)
A. Luo, F. Yang, K. Luo, X. Li, H. Fan and S. Liu: Learning Optical Flow with Adaptive Graph
Reasoning . AAAI 2022.
60
SSTMT++-tt-main [mv]
code
4.36 %
6.65 %
4.74 %
100.00 %
0.4 s
GPU @ 2.5 Ghz (Python)
F. Ferede and M. Balasubramanian: SSTM: Spatiotemporal Recurrent
Transformers for Multi-frame Optical Flow
Estimation . arXiv preprint arXiv:2304.14418 2023.
61
LSCA
4.49 %
6.03 %
4.75 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
62
HMAFlow
4.49 %
6.08 %
4.75 %
100.00 %
1 s
GPU (Python)
63
RAFT-A
code
4.54 %
5.99 %
4.78 %
100.00 %
0.7 s
GPU @ 2.5 Ghz (Python + C/C++)
D. Sun, D. Vlasic, C. Herrmann, V. Jampani, M. Krainin, H. Chang, R. Zabih, W. Freeman and C. Liu: AutoFlow: Learning a Better Training Set
for Optical Flow . CVPR 2021.
64
CRAFT
code
4.58 %
5.85 %
4.79 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
X. Sui, S. Li, X. Geng, Y. Wu, X. Xu, Y. Liu, R. Goh and H. Zhu: CRAFT: Cross-Attentional Flow Transformers
for Robust Optical Flow . CVPR 2022.
65
GMFlowNet
code
4.39 %
6.84 %
4.79 %
100.00 %
0.5 s
GPU @ 2.5 Ghz (Python)
S. Zhao, L. Zhao, Z. Zhang, E. Zhou and D. Metaxas: Global Matching with Overlapping Attention
for Optical Flow Estimation . CVPR 2022.
66
Scale-flow-ADF58
4.36 %
7.00 %
4.80 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (C/C++)
67
ce_skii
4.53 %
6.22 %
4.81 %
100.00 %
0.08 s
1 core @ 2.5 Ghz (C/C++)
68
SSTM++_ttt [mv]
4.45 %
6.71 %
4.83 %
100.00 %
0.3 s
GPU @ >3.5 Ghz (Python)
F. Ferede and M. Balasubramanian: SSTM: Spatiotemporal recurrent
transformers for multi-frame optical flow
estimation . Neurocomputing 2023.
69
llatst
4.51 %
6.57 %
4.85 %
100.00 %
2..4 s
1 core @ 2.5 Ghz (C/C++)
70
CE_SKFlow
4.55 %
6.39 %
4.85 %
100.00 %
0.08 s
1 core @ 2.5 Ghz (C/C++)
71
raft-sd
4.55 %
6.40 %
4.86 %
100.00 %
2.4 s
1 core @ 2.5 Ghz (C/C++)
72
MS_RAFT
4.58 %
6.38 %
4.88 %
100.00 %
0.3 s
GPU: Nvidia A100 (Python)
A. Jahedi, L. Mehl, M. Rivinius and A. Bruhn: Multi-Scale Raft: Combining Hierarchical Concepts for Learning-Based Optical Flow Estimation . IEEE International Conference on Image Processing (ICIP) 2022.
73
raft-aug
4.56 %
6.56 %
4.89 %
100.00 %
2.4 s
1 core @ 2.5 Ghz (C/C++)
74
DEQ-Flow-H
code
4.68 %
6.06 %
4.91 %
100.00 %
0.5 s
GPU @ 2.5 Ghz (Python)
S. Bai, Z. Geng, Y. Savani and Z. Kolter: Deep Equilibrium Optical Flow
Estimation . CVPR 2022.
75
MPI-Flow
code
4.40 %
7.48 %
4.91 %
100.00 %
1 s
1 core @ 2.5 Ghz (Python)
Y. Liang, J. Liu, D. Zhang and Y. Fu: Mpi-flow: Learning realistic optical flow
with multiplane images . Proceedings of the IEEE/CVF
International Conference on Computer Vision 2023.
76
LLA-Flow+GMA
4.57 %
6.68 %
4.92 %
100.00 %
0.24 s
1 core @ 2.5 Ghz (C/C++)
77
LLA-Flow+GMAv2
4.57 %
6.76 %
4.93 %
100.00 %
2.4 s
1 core @ 2.5 Ghz (C/C++)
78
MeFlow
4.59 %
6.75 %
4.95 %
100.00 %
0.12 s
1 core @ 2.5 Ghz (C/C++)
79
CSFlow
code
4.71 %
6.46 %
5.00 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
H. Shi, Y. Zhou, K. Yang, X. Yin and K. Wang: CSFlow: Learning Optical Flow via Cross
Strip Correlation for Autonomous Driving . arXiv preprint arXiv:2202.00909 2022.
80
SSTM_thes_[mv]
4.58 %
7.20 %
5.02 %
100.00 %
0.3 s
GPU @ 2.5 Ghz (Python)
F. Ferede: Multi-Frame Optical Flow Estimation Using Spatio-Temporal Transformers . 2022.
81
SSTM++_thes_[mv]
4.64 %
7.04 %
5.04 %
100.00 %
0.4 s
GPU @ 2.5 Ghz (Python)
F. Ferede: Multi-Frame Optical Flow Estimation Using Spatio-Temporal Transformers . 2022.
82
RAFT+AOIR
4.68 %
6.99 %
5.07 %
100.00 %
10 s
GPU @ 2.5 Ghz (Python + C/C++)
L. Mehl, C. Beschle, A. Barth and A. Bruhn: An Anisotropic Selection Scheme for Variational Optical Flow Methods with Order-Adaptive Regularisation . SSVM 2021.
83
RAFT
code
4.74 %
6.87 %
5.10 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
Z. Teed and J. Deng: RAFT: Recurrent All-Pairs Field Transforms
for Optical Flow . ECCV 2020.
84
LKCMA
4.85 %
6.65 %
5.15 %
100.00 %
0.2 s
GPU @ 2.5 Ghz (Python)
85
Scale-flow
code
5.24 %
5.71 %
5.32 %
100.00 %
0.8 s
GPU @ 2.5 Ghz (Python)
H. Ling, Q. Sun, Z. Ren, Y. Liu, H. Wang and Z. Wang: Scale-flow: Estimating 3D Motion from
Video . Proceedings of the 30th ACM
International Conference on Multimedia 2022.
86
RAFT-ADF
5.02 %
6.93 %
5.34 %
100.00 %
0.05 s
GPU @ 2.5 Ghz (Python + C/C++)
87
PRAFlow_RVC
5.08 %
7.21 %
5.43 %
100.00 %
0.5 s
GPU @ NVIDIA RTX 2080Ti (Python)
Z. Wan, Y. Mao and Y. Dai: PRAFlow_RVC: Pyramid Recurrent All-
Pairs Field Transforms for Optical Flow Estimation
in Robust Vision Challenge 2020 . 2020.
88
HCVNet
5.24 %
7.01 %
5.54 %
100.00 %
0.24 s
1 core @ 2.5 Ghz (C/C++)
89
RAFT-TF_RVC
5.32 %
6.75 %
5.56 %
100.00 %
0.7 s
GPU @ 2.5 Ghz (Python)
D. Sun, C. Herrmann, V. Jampani, M. Krainin, F. Cole, A. Stone, R. Jonschkowski, R. Zabih, W. Freeman and C. Liu: A TensorFlow implementation of RAFT . 2020.
90
MobileFlow
5.41 %
7.53 %
5.76 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (Python)
91
ACOSF
4.56 %
12.00 %
5.79 %
100.00 %
5 min
1 core @ 3.0 Ghz (Matlab + C/C++)
C. Li, H. Ma and Q. Liao: Two-Stage Adaptive Object Scene Flow Using
Hybrid CNN-CRF Model . International Conference on Pattern
Recognition (ICPR) 2020.
92
PPAC-HD3
code
5.78 %
7.48 %
6.06 %
100.00 %
0.19 s
NVIDIA GTX 1080 Ti
A. Wannenwetsch and S. Roth: Probabilistic Pixel-Adaptive Refinement Networks . CVPR 2020.
93
MaskFlownet
code
5.79 %
7.70 %
6.11 %
100.00 %
0.06 s
NVIDIA TITAN Xp
S. Zhao, Y. Sheng, Y. Dong, E. Chang and Y. Xu: MaskFlownet: Asymmetric Feature Matching
with Learnable Occlusion Mask . Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR) 2020.
94
RAFT+LCT-Flow
code
5.49 %
9.19 %
6.11 %
100.00 %
0.65 s
GPU @ 1.5 Ghz (Python + C/C++)
J. Chen: Motion Estimation with L0 norm
Regularization (Extended Version) . IEEE 7th International Conference on
Virtual Reality(ICVR) 2021.
95
RAPIDFlow
code
6.11 %
6.19 %
6.12 %
100.00 %
0.04 s
GPU @ 2.5 Ghz (Python)
H. Morimitsu, X. Zhu, R. Cesar-Jr, X. Ji and X. Yin: RAPIDFlow: {Recurrent Adaptable
Pyramids with Iterative Decoding} for Efficient
Optical Flow Estimation . ICRA 2024.
96
ISF
5.40 %
10.29 %
6.22 %
100.00 %
10 min
1 core @ 3 Ghz (C/C++)
A. Behl, O. Jafari, S. Mustikovela, H. Alhaija, C. Rother and A. Geiger: Bounding Boxes, Segmentations and Object
Coordinates: How Important is Recognition for 3D
Scene Flow Estimation in Autonomous Driving
Scenarios? . International Conference on Computer
Vision (ICCV) 2017.
97
VCN+LCV
code
5.75 %
8.80 %
6.25 %
100.00 %
0.26 s
1 core @ 2.5 Ghz (Python)
T. Xiao, J. Yuan, D. Sun, Q. Wang, X. Zhang, K. Xu and M. Yang: Learnable Cost Volume using the
Cayley Representation . Proceedings of the European
Conference on Computer Vision (ECCV) 2020.
98
RAFT+LCV
code
5.73 %
8.90 %
6.26 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (C/C++)
T. Xiao, J. Yuan, D. Sun, Q. Wang, X. Zhang, K. Xu and M. Yang: Learnable Cost Volume using the
Cayley Representation . Proceedings of the European
Conference on Computer Vision (ECCV) 2020.
99
PRichFlow
6.18 %
6.89 %
6.30 %
100.00 %
0.1 s
TITAN X MAXWELL
X. Wang, D. Zhu, J. Song, Y. Liu, J. Li and X. Zhang: Richer Aggregated Features for Optical Flow Estimation with Edge-aware Refinement . .
100
VCN
code
5.83 %
8.66 %
6.30 %
100.00 %
0.18 s
Titan X Pascal
G. Yang and D. Ramanan: Volumetric Correspondence Networks for
Optical Flow . NeurIPS 2019.
101
Stereo expansion
code
5.83 %
8.66 %
6.30 %
100.00 %
2 s
GPU @ 2.5 Ghz (Python)
G. Yang and D. Ramanan: Upgrading Optical Flow to 3D Scene Flow
through Optical Expansion . CVPR 2020.
102
Binary TTC
5.84 %
8.67 %
6.31 %
100.00 %
2 s
GPU @ 1.0 Ghz (Python)
A. Badki, O. Gallo, J. Kautz and P. Sen: Binary TTC: A Temporal Geofence for Autonomous
Navigation . The IEEE Conference on Computer Vision and Pattern
Recognition (CVPR) 2021.
103
MonoComb
5.84 %
8.67 %
6.31 %
100.00 %
0.58 s
RTX 2080 Ti
R. Schuster, C. Unger and D. Stricker: MonoComb: A Sparse-to-Dense Combination Approach for Monocular Scene Flow . ACM Computer Science in Cars Symposium (CSCS) 2020.
104
HD^3-Flow
code
6.05 %
9.02 %
6.55 %
100.00 %
0.10 s
NVIDIA Pascal Titan XP
Z. Yin, T. Darrell and F. Yu: Hierarchical Discrete Distribution Decomposition
for Match Density Estimation . CVPR 2019.
105
HOR-RAFT
6.22 %
8.41 %
6.59 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (Python)
106
PRSM
code
5.33 %
13.40 %
6.68 %
100.00 %
300 s
1 core @ 2.5 Ghz (C/C++)
C. Vogel, K. Schindler and S. Roth: 3D Scene Flow Estimation with a
Piecewise Rigid Scene Model . ijcv 2015.
107
MaskFlownet-S
code
6.53 %
8.21 %
6.81 %
100.00 %
0.03 s
NVIDIA TITAN Xp
S. Zhao, Y. Sheng, Y. Dong, E. Chang and Y. Xu: MaskFlownet: Asymmetric Feature Matching
with Learnable Occlusion Mask . Proceedings of the IEEE Conference on
Computer Vision and Pattern Recognition (CVPR) 2020.
108
ScopeFlow
code
6.72 %
7.36 %
6.82 %
100.00 %
-1 s
Nvidia GPU
A. Bar-Haim and L. Wolf: ScopeFlow: Dynamic Scene Scoping for
Optical Flow . The IEEE Conference on Computer
Vision and Pattern Recognition (CVPR) 2020.
109
SMURF
code
6.04 %
10.75 %
6.83 %
100.00 %
.2 s
1 core @ 2.5 Ghz (C/C++)
A. Stone, D. Maurer, A. Ayvaci, A. Angelova and R. Jonschkowski: SMURF: Self-Teaching Multi-Frame
Unsupervised RAFT With Full-Image Warping . Proceedings of the IEEE/CVF
Conference on Computer Vision and Pattern
Recognition (CVPR) 2021.
110
OSF+TC
5.76 %
13.31 %
7.02 %
100.00 %
50 min
1 core @ 2.5 Ghz (C/C++)
M. Neoral and J. Šochman: Object Scene Flow with Temporal
Consistency . 22nd Computer Vision Winter
Workshop (CVWW) 2017.
111
DPCTF-F
7.22 %
6.47 %
7.09 %
100.00 %
0.07 s
GPU @ 2.5 Ghz (C/C++)
Y. Deng, J. Xiao, S. Zhou and J. Feng: Detail Preserving Coarse-to-Fine Matching
for Stereo Matching and Optical Flow . IEEE Transactions on Image Processing 2021.
112
SSF
5.63 %
14.71 %
7.14 %
100.00 %
5 min
1 core @ 2.5 Ghz (Matlab + C/C++)
Z. Ren, D. Sun, J. Kautz and E. Sudderth: Cascaded Scene Flow Prediction using
Semantic Segmentation . International Conference on 3D Vision
(3DV) 2017.
113
MFF
7.15 %
7.25 %
7.17 %
100.00 %
0.05 s
NVIDIA Pascal Titan X (Python)
Z. Ren, O. Gallo, D. Sun, M. Yang, E. Sudderth and J. Kautz: A Fusion Approach for Multi-Frame
Optical
Flow Estimation . IEEE Winter Conference on
Applications of Computer Vision 2019.
114
LiteFlowNet3-S
code
7.27 %
6.96 %
7.22 %
100.00 %
0.07s
GTX 1080 (slower than Titan X Pascal)
T. Hui and C. Loy: LiteFlowNet3: Resolving Correspondence
Ambiguity for More Accurate Optical Flow
Estimation . European Conference on Computer Vision
(ECCV) 2020.
115
PMC-PWC
code
7.27 %
6.94 %
7.22 %
100.00 %
TBD s
GPU @ 2.5 Ghz (Python)
C. Zhang, C. Feng, Z. Chen, W. Hu and M. Li: Parallel multiscale context-based edge-
preserving optical flow estimation with occlusion
detection . Signal Processing: Image
Communication 2022.
116
SwiftFlow
6.85 %
9.11 %
7.23 %
100.00 %
0.03 s
GPU @ 2.5 Ghz (Python)
H. Wang, Y. Liu, H. Huang, Y. Pan, W. Yu, J. Jiang, D. Lyu, M. Bocus, M. Liu, I. Pitas and others: ATG-PVD: Ticketing parking
violations on a drone . European Conference on Computer
Vision 2020.
117
LiteFlowNet3
code
7.26 %
7.75 %
7.34 %
100.00 %
0.07s
GTX 1080 (slower than Titan X Pascal)
T. Hui and C. Loy: LiteFlowNet3: Resolving Correspondence
Ambiguity for More Accurate Optical Flow
Estimation . European Conference on Computer Vision
(ECCV) 2020.
118
UFD-PRiME
5.85 %
14.91 %
7.36 %
100.00 %
0.56 s
GPU @ 2.5 Ghz (Python)
119
M2Flow
6.16 %
13.44 %
7.37 %
100.00 %
0.09 s
1 core @ 2.5 Ghz (Python)
120
OSF 2018
code
5.38 %
17.61 %
7.41 %
100.00 %
390 s
1 core @ 2.5 Ghz (Matlab + C/C++)
M. Menze, C. Heipke and A. Geiger: Object Scene Flow . ISPRS Journal of Photogrammetry and Remote Sensing (JPRS) 2018.
121
LiteFlowNet2
code
7.62 %
7.64 %
7.62 %
100.00 %
0.0486 s
GTX 1080 (slower than Titan X Pascal)
T. Hui, X. Tang and C. Loy: A Lightweight Optical Flow CNN -
Revisiting Data
Fidelity and Regularization . TPAMI 2020.
122
UFD-PRiME
5.96 %
15.96 %
7.63 %
100.00 %
0.55 s
GPU @ 2.5 Ghz (Python)
123
SENSE
code
7.30 %
9.33 %
7.64 %
100.00 %
0.32s
GPU, GTX 2080Ti
H. Jiang, D. Sun, V. Jampani, Z. Lv, E. Learned-Miller and J. Kautz: SENSE: A Shared Encoder Network for Scene-Flow
Estimation . The IEEE International Conference on Computer
Vision (ICCV) 2019.
124
IRR-PWC
code
7.68 %
7.52 %
7.65 %
100.00 %
0.18 s
NVIDIA GTX 1080 Ti
J. Hur and S. Roth: Iterative Residual Refinement for
Joint Optical Flow and Occlusion Estimation . CVPR 2019.
125
STaRFlow
code
7.51 %
8.35 %
7.65 %
100.00 %
0.24 s
GPU @ 2.0 Ghz (Python)
P. Godet, A. Boulch, A. Plyer and G. Besnerais: STaRFlow: A SpatioTemporal Recurrent Cell for Lightweight Multi-Frame Optical Flow Estimation . ICPR 2020.
126
DTF_SENSE
7.31 %
9.48 %
7.67 %
100.00 %
0.76 s
1 core @ 2.5 Ghz (C/C++)
R. Schuster, C. Unger and D. Stricker: A Deep Temporal Fusion Framework for Scene Flow Using a Learnable Motion Model and Occlusions . IEEE Winter Conference on Applications of Computer Vision (WACV) 2021.
127
PWC-Net+
code
7.69 %
7.88 %
7.72 %
100.00 %
0.03 s
NVIDIA Pascal Titan X
D. Sun, X. Yang, M. Liu and J. Kautz: Models Matter, So Does Training: An
Empirical Study of CNNs for Optical Flow
Estimation . arXiv preprint arXiv:1809.05571 2018.
128
raft-s
7.30 %
10.02 %
7.76 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (Python)
129
UnSAMFlow
6.40 %
14.98 %
7.83 %
100.00 %
0.03 s
GPU @ 2.5 Ghz (Python)
130
OSF
code
5.62 %
18.92 %
7.83 %
100.00 %
50 min
1 core @ 2.5 Ghz (C/C++)
M. Menze and A. Geiger: Object Scene Flow for Autonomous Vehicles . Conference on Computer Vision and Pattern Recognition (CVPR) 2015.
131
Separable-Sim2real
7.30 %
11.01 %
7.92 %
100.00 %
0.25 s
GPU
F. Zhang, O. Woodford, V. Prisacariu and P. Torr: Separable Flow: Learning Motion Cost
Volumes for Optical Flow Estimation . Proceedings of the IEEE/CVF
International Conference on Computer Vision 2021.
132
LSM_FLOW_RVC
code
7.33 %
13.06 %
8.28 %
100.00 %
0.2 s
1 core @ 2.5 Ghz (C/C++)
C. Tang, L. Yuan and P. Tan: LSM: Learning Subspace Minimization for Low-Level Vision . IEEE/CVF Conference on Computer Vision and Pattern
Recognition (CVPR) 2020.
133
AL-OF_r0.2
code
7.25 %
13.53 %
8.30 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (Python)
S. Yuan, X. Sun, H. Kim, S. Yu and C. Tomasi: Optical Flow Training Under Limited Label Budget via Active
Learning . ECCV 2022.
134
IRR-PWC_RVC
code
7.61 %
12.22 %
8.38 %
100.00 %
0.18 s
NVIDIA GTX 1080 Ti
J. Hur and S. Roth: Iterative Residual Refinement for
Joint Optical Flow and Occlusion Estimation . CVPR 2019.
135
SemARFlow
code
7.48 %
12.91 %
8.38 %
100.00 %
0.0168s
GPU @ 2.5 Ghz (Python)
S. Yuan, S. Yu, H. Kim and C. Tomasi: SemARFlow: Injecting Semantics into
Unsupervised Optical Flow Estimation for Autonomous
Driving . ICCV 2023.
136
SelFlow
7.61 %
12.48 %
8.42 %
100.00 %
0.09 s
GPU @ 2.5 Ghz (Python)
P. Liu, M. Lyu, I. King and J. Xu: SelFlow: Self-Supervised Learning of Optical
Flow . CVPR 2019.
137
MDFlow
8.14 %
12.80 %
8.91 %
100.00 %
0.03 s
NVIDIA GTX 1080 Ti
L. Kong and J. Yang: MDFlow: Unsupervised Optical Flow Learning
by Reliable Mutual Knowledge Distillation . IEEE Transactions on Circuits and Systems
for Video Technology 2022.
138
GMFlow
code
9.67 %
7.57 %
9.32 %
100.00 %
0.071 s
A100 GPU (Python)
H. Xu, J. Zhang, J. Cai, H. Rezatofighi and D. Tao: GMFlow: Learning Optical Flow via Global
Matching . Proceedings of the IEEE/CVF Conference on
Computer Vision and Pattern Recognition 2022.
139
FDFlowNet
9.31 %
9.71 %
9.38 %
100.00 %
0.02 s
NVIDIA GTX 1080 Ti
L. Kong and J. Yang: FDFlowNet: Fast Optical Flow Estimation
using a Deep Lightweight Network . IEEE International Conference on
Image Processing (ICIP) 2020.
140
LiteFlowNet
code
9.66 %
7.99 %
9.38 %
100.00 %
0.0885 s
GTX 1080 (slower than Titan X Pascal)
T. Hui, X. Tang and C. Loy: LiteFlowNet: A Lightweight
Convolutional Neural Network for Optical Flow
Estimation . Proceedings of IEEE Conference
on Computer Vision and Pattern Recognition
(CVPR) 2018.
141
PWC-Net
code
9.66 %
9.31 %
9.60 %
100.00 %
0.03 s
NVIDIA Pascal Titan X
D. Sun, X. Yang, M. Liu and J. Kautz: PWC-Net: CNNs for Optical
Flow Using Pyramid, Warping, and Cost Volume . CVPR 2018.
142
ScaleFlow++_SAG
9.78 %
9.50 %
9.73 %
100.00 %
0.05 s
GPU @ 2.5 Ghz (Python + C/C++)
143
NeuFlow
code
9.49 %
11.82 %
9.88 %
100.00 %
0.01 s
GPU @ 2.5 Ghz (C/C++)
ERROR: Wrong syntax in BIBTEX file.
144
ContinualFlow_ROB
8.54 %
17.48 %
10.03 %
100.00 %
0.15 s
GPU - NVidia 1080Ti
M. Neoral, J. Šochman and J. Matas: Continual Occlusions and Optical Flow
Estimation . 14th Asian Conference on Computer
Vision (ACCV) 2018.
145
VCN_RVC
code
8.53 %
18.30 %
10.15 %
100.00 %
0.36 s
GPU @ 2.5 Ghz (Python)
G. Yang and D. Ramanan: Volumetric Correspondence Networks for
Optical Flow . NeurIPS 2019.
146
NccFLow
8.81 %
17.36 %
10.24 %
100.00 %
0.04 s
1 core @ 2.5 Ghz (C/C++)
G. Wang, S. Ren and H. Wang: NccFlow: Unsupervised Learning of Optical
Flow With Non-occlusion from Geometry . arXiv preprint arXiv:2107.03610 2021.
147
MirrorFlow
code
8.93 %
17.07 %
10.29 %
100.00 %
11 min
4 core @ 2.2 Ghz (C/C++)
J. Hur and S. Roth: MirrorFlow: Exploiting Symmetries
in Joint Optical Flow and Occlusion Estimation . ICCV 2017.
148
CoT-AMFlow
10.02 %
11.95 %
10.34 %
100.00 %
0.06 s
GPU @ 2.5 Ghz (Python)
H. Wang, R. Fan and M. Liu: CoT-AMFlow: Adaptive Modulation Network
with Co-Teaching Strategy for Unsupervised Optical
Flow Estimation . Conference on Robot Learning (CoRL) 2020.
149
DWARF
9.80 %
13.37 %
10.39 %
100.00 %
0.14s - 1.43s
TitanXP - JetsonTX2
F. Aleotti, M. Poggi, F. Tosi and S. Mattoccia: Learning end-to-end scene flow by
distilling single tasks knowledge . Thirty-Fourth AAAI Conference on
Artificial Intelligence (AAAI-20) 2020.
150
FlowNet2
code
10.75 %
8.75 %
10.41 %
100.00 %
0.1 s
GPU @ 2.5 Ghz (C/C++)
E. Ilg, N. Mayer, T. Saikia, M. Keuper, A. Dosovitskiy and T. Brox: FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks . IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2017.
151
SDF
8.61 %
23.01 %
11.01 %
100.00 %
TBA
1 core @ 2.5 Ghz (C/C++)
M. Bai*, W. Luo*, K. Kundu and R. Urtasun: Exploiting Semantic Information and Deep
Matching for Optical Flow . ECCV 2016.
152
GS58_ScaleRES
11.09 %
10.77 %
11.04 %
100.00 %
0.1 s
GPU @ 2.5 Ghz (Python + C/C++)
153
Flow2Stereo
9.99 %
16.67 %
11.10 %
100.00 %
0.05 s
GPU @ 2.5 Ghz (Python)
P. Liu, I. King, M. Lyu and J. Xu: Flow2Stereo: Effective Self-Supervised
Learning of Optical Flow and Stereo Matching . CVPR 2020.
154
UnFlow
code
10.15 %
15.93 %
11.11 %
100.00 %
0.12 s
GPU @ 1.5 Ghz (Python + C/C++)
S. Meister, J. Hur and S. Roth: UnFlow: Unsupervised Learning of Optical
Flow with a Bidirectional Census Loss . AAAI 2018.
155
UFlow
code
9.78 %
17.87 %
11.13 %
100.00 %
0.04 s
1 core @ 2.5 Ghz (C/C++)
R. Jonschkowski, A. Stone, J. Barron, A. Gordon, K. Konolige and A. Angelova: What Matters in Unsupervised Optical
Flow . ECCV 2020.
156
ADF_RESScaleFlow
11.28 %
10.66 %
11.18 %
100.00 %
0.1 s
GPU @ 2.5 Ghz (Python + C/C++)
157
FastFlowNet
code
11.20 %
11.30 %
11.22 %
100.00 %
0.01 s
NVIDIA GTX 1080 Ti
L. Kong, C. Shen and J. Yang: FastFlowNet: A Lightweight Network for Fast
Optical Flow Estimation . 2021 IEEE International Conference on
Robotics and Automation (ICRA) 2021.
158
FSF+MS
8.48 %
25.43 %
11.30 %
100.00 %
2.7 s
4 cores @ 3.5 Ghz (C/C++)
T. Taniai, S. Sinha and Y. Sato: Fast Multi-frame Stereo Scene Flow
with Motion Segmentation . IEEE Conference on Computer Vision
and Pattern Recognition (CVPR 2017) 2017.
159
GS58_Scale
11.22 %
11.76 %
11.31 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (Python)
160
MDFlow-Fast
10.75 %
14.81 %
11.43 %
100.00 %
0.01 s
NVIDIA GTX 1080 Ti
L. Kong and J. Yang: MDFlow: Unsupervised Optical Flow Learning
by Reliable Mutual Knowledge Distillation . IEEE Transactions on Circuits and Systems
for Video Technology 2022.
161
CNNF+PMBP
10.08 %
18.56 %
11.49 %
100.00 %
45 min
1 cores @ 3.5 Ghz (C/C++)
F. Zhang and B. Wah: Fundamental Principles on Learning New
Features for Effective Dense Matching . IEEE Transactions on Image
Processing 2018.
162
MaxFlow
code
12.02 %
9.05 %
11.52 %
100.00 %
1 s
GPU @ >3.5 Ghz (Python)
163
PWC-Net_RVC
code
11.22 %
13.69 %
11.63 %
100.00 %
0.03 s
NVIDIA Pascal Titan X
D. Sun, X. Yang, M. Liu and J. Kautz: PWC-Net: CNNs for Optical Flow
Using Pyramid, Warping, and Cost Volume . CVPR 2018.
164
SFF++
10.63 %
17.48 %
11.77 %
100.00 %
78 s
4 cores @ 3.5 Ghz (C/C++)
R. Schuster, O. Wasenmüller, C. Unger, G. Kuschk and D. Stricker: SceneFlowFields++: Multi-frame Matching, Visibility Prediction, and Robust Interpolation for Scene Flow Estimation . International Journal of Computer Vision (IJCV) 2019.
165
SfM-PM
9.66 %
22.73 %
11.83 %
100.00 %
69 s
3 cores @ 3.6 Ghz (C/C++)
D. Maurer, N. Marniok, B. Goldluecke and A. Bruhn: Structure-from-Motion-Aware PatchMatch for Adaptive Optical Flow Estimation . ECCV 2018.
166
EMR-MSF
9.86 %
22.27 %
11.93 %
100.00 %
0.25 s
GPU @ 2.5 Ghz (Python)
Z. Jiang and M. Okutomi: EMR-MSF: Self-Supervised Recurrent
Monocular Scene Flow Exploiting Ego-Motion
Rigidity . Proceedings of the IEEE/CVF
International Conference on Computer Vision 2023.
167
Self-SuperFlow-ft
10.65 %
19.44 %
12.12 %
100.00 %
0.13 s
GTX 1080 Ti
K. Bendig, R. Schuster and D. Stricker: Self-SuperFlow: Self-supervised Scene Flow Prediction in
Stereo Sequences . International Conference on Image Processing (ICIP) 2022.
168
MR-Flow
code
10.13 %
22.51 %
12.19 %
100.00 %
8 min
1 core @ 2.5 Ghz (Python + C/C++)
J. Wulff, L. Sevilla-Lara and M. Black: Optical Flow in Mostly Rigid Scenes . IEEE Conf. on Computer Vision and Pattern Recognition (CVPR) 2017.
169
DTF_PWOC
10.78 %
19.99 %
12.31 %
100.00 %
0.38 s
RTX 2080 Ti
R. Schuster, C. Unger and D. Stricker: A Deep Temporal Fusion Framework for Scene Flow Using a Learnable Motion Model and Occlusions . IEEE Winter Conference on Applications of Computer Vision (WACV) 2021.
170
Mono-SF
11.40 %
19.64 %
12.77 %
100.00 %
41 s
1 core @ 3.5 Ghz (Matlab + C/C++)
F. Brickwedde, S. Abraham and R. Mester: Mono-SF: Multi-View Geometry meets Single-View Depth for Monocular Scene Flow Estimation of Dynamic Traffic Scenes . Proc. of International Conference on Computer Vision (ICCV) 2019.
171
SceneFFields
10.58 %
24.41 %
12.88 %
100.00 %
65 s
4 cores @ 3.7 Ghz (C/C++)
R. Schuster, O. Wasenmüller, G. Kuschk, C. Bailer and D. Stricker: SceneFlowFields: Dense Interpolation of Sparse Scene Flow Correspondences . IEEE Winter Conference on Applications of Computer Vision (WACV) 2018.
172
CSF
10.40 %
25.78 %
12.96 %
100.00 %
80 s
1 core @ 2.5 Ghz (C/C++)
Z. Lv, C. Beall, P. Alcantarilla, F. Li, Z. Kira and F. Dellaert: A Continuous Optimization Approach for
Efficient and Accurate Scene Flow . European Conf. on Computer Vision
(ECCV) 2016.
173
PWOC-3D
code
12.40 %
15.78 %
12.96 %
100.00 %
0.13 s
GTX 1080 Ti
R. Saxena, R. Schuster, O. Wasenmüller and D. Stricker: PWOC-3D: Deep Occlusion-Aware End-to-End Scene Flow Estimation . Intelligent Vehicles Symposium (IV) 2019.
174
Multi-Mono-SF-ft
code
12.41 %
18.20 %
13.37 %
100.00 %
0.06 s
NVIDIA GTX 1080 Ti
J. Hur and S. Roth: Self-Supervised Multi-Frame
Monocular Scene Flow . CVPR 2021.
175
UnsupSimFlow
code
12.60 %
17.27 %
13.38 %
100.00 %
0.03 s
8 cores @ 3.0 Ghz (Python + C/C++)
W. Im, T. Kim and S. Yoon: Unsupervised Learning of Optical Flow
with Deep Feature Similarity . The European Conference on Computer
Vision (ECCV) 2020.
176
Self-scale-flow-nerf
13.08 %
15.45 %
13.47 %
100.00 %
0.2 s
1 core @ 2.5 Ghz (C/C++)
177
Nerf-self
13.35 %
14.65 %
13.57 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (C/C++)
178
PR-Sceneflow
code
11.73 %
24.33 %
13.83 %
100.00 %
150 s
4 core @ 3.0 Ghz (Matlab + C/C++)
C. Vogel, K. Schindler and S. Roth: Piecewise Rigid Scene Flow . ICCV 2013.
179
DDFlow+LCV
12.98 %
19.83 %
14.12 %
100.00 %
0.1 s
GPU @ 2.5 Ghz (Python)
T. Xiao, J. Yuan, D. Sun, Q. Wang, X. Zhang, K. Xu and M. Yang: Learnable Cost Volume using the
Cayley Representation . Proceedings of the European
Conference on Computer Vision (ECCV) 2020.
180
SelFlow
12.68 %
21.74 %
14.19 %
100.00 %
0.09 s
GPU @ 2.5 Ghz (Python)
P. Liu, M. Lyu, I. King and J. Xu: SelFlow: Self-Supervised Learning of Optical
Flow . CVPR 2019.
181
DDFlow
13.08 %
20.40 %
14.29 %
100.00 %
0.06 s
GPU @ >3.5 Ghz (Python + C/C++)
P. Liu, I. King, M. Lyu and J. Xu: DDFlow: Learning Optical Flow with Unlabeled
Data Distillation . AAAI 2019.
182
DCFlow
code
13.10 %
23.70 %
14.86 %
100.00 %
8.6 s
GPU @ 3.0 Ghz (Matlab + C/C++)
J. Xu, R. Ranftl and V. Koltun: Accurate Optical Flow via Direct
Cost Volume Processing . CVPR 2017.
183
ProFlow
13.86 %
20.91 %
15.04 %
100.00 %
112 s
GPU+CPU @ 3.6 Ghz (Python + C/C++)
D. Maurer and A. Bruhn: ProFlow: Learning to Predict Optical Flow . BMVC 2018.
184
FlowFields++
code
14.82 %
17.77 %
15.31 %
100.00 %
29 s
1 core @ 3.5 Ghz (C/C++)
R. Schuster, C. Bailer, O. Wasenmüller and D. Stricker: FlowFields++: Accurate Optical Flow Correspondences Meet Robust Interpolation . International Conference on Image Processing (ICIP) 2018.
185
GS_ScaleFlow
15.29 %
15.64 %
15.35 %
100.00 %
0.1 s
GPU @ 2.5 Ghz (Python)
186
ProFlow_ROB
14.15 %
21.82 %
15.42 %
100.00 %
112 s
GPU+CPU @ 3.6 Ghz (Python + C/C++)
D. Maurer and A. Bruhn: ProFlow: Learning to Predict Optical Flow . BMVC 2018.
187
Self-Mono-SF-ft
code
15.51 %
17.96 %
15.91 %
100.00 %
0.09 s
NVIDIA GTX 1080 Ti
J. Hur and S. Roth: Self-Supervised Monocular Scene
Flow Estimation . CVPR 2020.
188
FF++_ROB
15.32 %
19.27 %
15.97 %
100.00 %
29 s
1 core @ 3.5 Ghz (C/C++)
R. Schuster, C. Bailer, O. Wasenmüller and D. Stricker: FlowFields++: Accurate Optical Flow Correspondences Meet Robust Interpolation . International Conference on Image Processing (ICIP) 2018.
189
SOF
code
14.63 %
22.83 %
15.99 %
100.00 %
6 min
1 core @ 2.5 Ghz (Matlab)
L. Sevilla-Lara, D. Sun, V. Jampani and M. Black: Optical Flow with Semantic Segmentation and
Localized Layers . CVPR 2016.
190
GMFlow+ADF58
15.70 %
18.41 %
16.15 %
100.00 %
0.1 s
1 core @ 2.5 Ghz (Python + C/C++)
191
DIP-Flow-DF
14.93 %
23.37 %
16.33 %
100.00 %
104s
2 cores @ 3.6 Ghz (C/C++)
D. Maurer, M. Stoll and A. Bruhn: Directional Priors for Multi-Frame Optical Flow . BMVC 2018.
192
JFS
15.90 %
19.31 %
16.47 %
100.00 %
13 min
1 core @ 3.2 Ghz (C/C++)
J. Hur and S. Roth: Joint Optical Flow and Temporally
Consistent Semantic Segmentation . ECCV Workshops 2016.
193
DF+OIR
15.11 %
23.45 %
16.50 %
100.00 %
3 min
1 core @ 3.5 Ghz (Matlab + C/C++)
D. Maurer, M. Stoll and A. Bruhn: Order-Adaptive and Illumination Aware Variational Optical Flow Refinement . BMVC 2017.
194
SPS+FF++
code
15.91 %
20.27 %
16.64 %
100.00 %
36 s
1 core @ 3.5 Ghz (C/C++)
R. Schuster, O. Wasenmüller and D. Stricker: Dense Scene Flow from Stereo Disparity and Optical Flow . ACM Computer Science in Cars Symposium (CSCS) 2018.
195
DIP-Flow-CPM
15.57 %
23.84 %
16.95 %
100.00 %
52 s
2 core @ 3.6 Ghz (C/C++)
D. Maurer, M. Stoll and A. Bruhn: Directional Priors for Multi-Frame Optical Flow . BMVC 2018.
196
ImpPB+SPCI
code
17.25 %
20.44 %
17.78 %
100.00 %
60 s
GPU @ 2.5 Ghz (Python)
T. Schuster, L. Wolf and D. Gadot: Optical Flow Requires Multiple Strategies
(but only one network) . CVPR 2017.
197
Anonymous
17.91 %
18.08 %
17.93 %
100.00 %
0.1 s
GPU @ 2.5 Ghz (Python)
198
PCOF-LDOF
14.34 %
38.32 %
18.33 %
100.00 %
50 s
1 core @ 3.0 Ghz (C/C++)
M. Derome, A. Plyer, M. Sanfourche and G. Le Besnerais: A Prediction-Correction Approach for Real-Time Optical Flow Computation Using Stereo . German Conference on Pattern Recognition 2016.
199
FlowFieldCNN
18.33 %
20.42 %
18.68 %
100.00 %
23 s
GPU/CPU 4 core @ 3.5 Ghz (C/C++)
C. Bailer, K. Varanasi and D. Stricker: CNN-based Patch Matching for Optical Flow with Thresholded Hinge Embedding Loss . CVPR 2017.
200
RicFlow
18.73 %
19.09 %
18.79 %
100.00 %
5 s
1 core @ 3.5 Ghz (C/C++)
Y. Hu, Y. Li and R. Song: Robust Interpolation of Correspondences for
Large Displacement Optical Flow . CVPR 2017.
201
HCSH
18.05 %
26.23 %
19.41 %
100.00 %
3.5 s
1 core @ 3.0 Ghz (C/C++)
J. Fan, Y. Wang and L. Guo: Hierarchical coherency sensitive hashing and interpolation with
RANSAC for large displacement optical flow . Computer Vision and Image Understanding 2018.
202
OmegaNet
17.43 %
29.69 %
19.47 %
100.00 %
0.01 s
GPU @ 1.5 Ghz (Python)
F. Tosi, F. Aleotti, P. Ramirez, M. Poggi, S. Salti, L. Di Stefano and S. Mattoccia: Distilled semantics for comprehensive scene understanding from videos . Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2020.
203
UJG
code
18.57 %
24.02 %
19.48 %
100.00 %
0.03 s
GPU @ 2.5 Ghz (Python)
J. Li, J. Zhao, S. Song and T. Feng: Unsupervised Joint Learning of Depth,
Optical Flow, Ego-motion from Video . arXiv preprint arXiv:2105.14520 2021.
204
Multi-Mono-SF
code
18.13 %
26.59 %
19.54 %
100.00 %
0.06 s
NVIDIA GTX 1080 Ti
J. Hur and S. Roth: Self-Supervised Multi-Frame
Monocular Scene Flow . CVPR 2021.
205
PGM-G
18.90 %
23.43 %
19.66 %
100.00 %
5.05 s
1 core @ 3.1 Ghz (C/C++)
Y. Li: Pyramidal Gradient Matching for
Optical Flow Estimation . CoRR 2017.
206
FlowFields+
19.51 %
21.26 %
19.80 %
100.00 %
28s
1 core @ 3.5 Ghz (C/C++)
C. Bailer, B. Taetz and D. Stricker: Flow Fields: Dense Correspondence Fields for Highly Accurate Large Displacement Optical Flow Estimation . .
207
EPC++ (stereo)
19.24 %
26.93 %
20.52 %
100.00 %
0.05 s
GPU @ 2.5 Ghz (Python)
C. Luo, Z. Yang, P. Wang, Y. Wang, W. Xu, R. Nevatia and A. Yuille: Every Pixel Counts ++: Joint Learning of
Geometry and Motion with 3D Holistic Understanding . IEEE transactions on pattern analysis and
machine intelligence 2019.
208
PatchBatch
code
19.98 %
26.50 %
21.07 %
100.00 %
50 s
GPU @ 2.5 Ghz (Python)
D. Gadot and L. Wolf: PatchBatch: a Batch Augmented Loss for
Optical Flow . The IEEE Conference on Computer Vision
and Pattern Recognition (CVPR) 2016.
209
DDF
code
20.36 %
25.19 %
21.17 %
100.00 %
~1 min
GPU @ 2.5 Ghz (C/C++)
F. G\"uney and A. Geiger: Deep Discrete Flow . Asian Conference on Computer Vision
(ACCV) 2016.
210
SODA-Flow
20.01 %
29.14 %
21.53 %
100.00 %
96 s
2 cores @ 3.5 Ghz (C/C++)
D. Maurer, M. Stoll, S. Volz, P. Gairing and A. Bruhn: A Comparison of Isotropic and Anisotropic Second Order Regularisers for Optical Flow . SSVM 2017.
211
DiscreteFlow
code
21.53 %
21.76 %
21.57 %
100.00 %
3 min
1 core @ 2.5 Ghz (Matlab + C/C++)
M. Menze, C. Heipke and A. Geiger: Discrete Optimization for Optical Flow . German Conference on Pattern
Recognition (GCPR) 2015.
212
SGM+SF
20.91 %
25.50 %
21.67 %
100.00 %
45 min
16 core @ 3.2 Ghz (C/C++)
H. Hirschmüller: Stereo Processing by Semiglobal Matching
and Mutual Information . PAMI 2008. M. Hornacek, A. Fitzgibbon and C. Rother: SphereFlow: 6
DoF Scene Flow from RGB-D Pairs . CVPR 2014.
213
OAR-Flow
20.62 %
27.67 %
21.79 %
100.00 %
100 s
2 cores @ 3.5 Ghz (C/C++)
D. Maurer, M. Stoll and A. Bruhn: Order-Adaptive Regularisation for Variational Optical Flow: Global, Local and in Between . SSVM 2017.
214
CPM-Flow
code
22.32 %
22.81 %
22.40 %
100.00 %
4.2 s
1 core @ 3.5 Ghz (C/C++)
Y. Hu, R. Song and Y. Li: Efficient Coarse-to-Fine PatchMatch for
Large
Displacement Optical Flow . CVPR 2016.
215
PCOF + ACTF
14.89 %
60.15 %
22.43 %
100.00 %
0.08 s
GPU @ 2.0 Ghz (C/C++)
M. Derome, A. Plyer, M. Sanfourche and G. Le Besnerais: A Prediction-Correction Approach for Real-Time Optical Flow Computation Using Stereo . German Conference on Pattern Recognition 2016.
216
SegFlow(d0=3)
22.21 %
23.72 %
22.46 %
100.00 %
6.6 s
1 core @ >3.5 Ghz (C/C++)
J. Chen, Z. Cai, J. Lai and X. Xie: Efficient Segmentation-based PatchMatch for
Large displacement Optical
Flow Estimation . IEEE TCSVT 2018.
217
IntrpNt-df
code
22.15 %
26.03 %
22.80 %
100.00 %
3 min
GPU @ 2.5 Ghz (Python)
S. Zweig and L. Wolf: InterpoNet, a Brain Inspired Neural
Network for Optical Flow Dense Interpolation . The IEEE Conference on Computer
Vision and Pattern Recognition (CVPR) 2017.
218
SGM&FlowFie+
22.83 %
22.75 %
22.82 %
81.24 %
29 s
1 core @ 3.5 Ghz (C/C++)
R. Schuster, C. Bailer, O. Wasenmüller and D. Stricker: Combining Stereo Disparity and Optical Flow for Basic Scene Flow . Commercial Vehicle Technology Symposium (CVTS) 2018.
219
Back2FutureFlow(UFO)
code
22.67 %
24.27 %
22.94 %
100.00 %
0.12 s
GPU @ 2.5 Ghz (LUA/Torch)
J. Janai, F. Güney, A. Ranjan, M. Black and A. Geiger: Unsupervised Learning of Multi-Frame Optical Flow with Occlusions . Proc. of the European Conf. on Computer Vision (ECCV) 2018.
220
MotionSLIC
code
14.86 %
64.44 %
23.11 %
100.00 %
30 s
4 cores @ 2.5 Ghz (C/C++)
K. Yamaguchi, D. McAllester and R. Urtasun: Robust Monocular Epipolar Flow Estimation . CVPR 2013.
221
IntrpNt-cpm
code
22.51 %
26.54 %
23.18 %
100.00 %
5.6 s
GPU @ 2.5 Ghz (Python)
S. Zweig and L. Wolf: InterpoNet, a Brain Inspired Neural
Network for Optical Flow Dense Interpolation . The IEEE Conference on Computer
Vision and Pattern Recognition (CVPR) 2017.
222
FullFlow
23.09 %
24.79 %
23.37 %
100.00 %
4 min
4 cores @ >3.5 Ghz (Matlab and C++)
Q. Chen and V. Koltun: Full Flow: Optical Flow Estimation By
Global Optimization over Regular Grids . CVPR 2016.
223
HiLM
code
23.73 %
21.79 %
23.41 %
100.00 %
8 sec
P6000 (C/C++)
M. Fathy, Q. Tran, M. Zia, P. Vernaza and M. Chandraker: Hierarchical Metric Learning and Matching for 2D and 3D Geometric Correspondences . European Conference on Computer Vision (ECCV) 2018.
224
Self-Mono-SF
code
23.26 %
24.93 %
23.54 %
100.00 %
0.09 s
NVIDIA GTX 1080 Ti
J. Hur and S. Roth: Self-Supervised Monocular Scene
Flow Estimation . CVPR 2020.
225
Self-SuperFlow
22.70 %
28.55 %
23.67 %
100.00 %
0.13 s
GTX 1080 Ti
K. Bendig, R. Schuster and D. Stricker: Self-SuperFlow: Self-supervised Scene Flow Prediction in
Stereo Sequences . International Conference on Image Processing (ICIP) 2022.
226
IntrpNt-dm
code
23.46 %
26.27 %
23.93 %
100.00 %
15 s
GPU @ 2.5 Ghz (Python)
S. Zweig and L. Wolf: InterpoNet, a Brain Inspired Neural
Network for Optical Flow Dense Interpolation . The IEEE Conference on Computer
Vision and Pattern Recognition (CVPR) 2017.
227
SPM-BP
24.06 %
24.97 %
24.21 %
100.00 %
10 s
2 cores @ 2.5 Ghz (C/C++)
Y. Li, D. Min, M. Brown, M. Do and J. Lu: SPM-BP: Sped-up PatchMatch Belief Propagation for
Continuous MRFs . Proceedings of the IEEE International
Conference on Computer Vision 2015.
228
PPM
code
25.87 %
23.67 %
25.50 %
100.00 %
17.3 s
1 core @ 2.5 Ghz (C/Chttps://github.c++)
F. Kuang: PatchMatch algorithms for motion
estimation and stereo reconstruction . 2017.
229
3DFlow
25.56 %
29.33 %
26.19 %
100.00 %
448s
Matlab with embedded C++ code
J. Chen, Z. Cai, J. Lai and X. Xie: A Filtering Based Framework for Optical
Flow Estimation . IEEE TCSVT 2018.
230
EpicFlow
code
25.81 %
28.69 %
26.29 %
100.00 %
15 s
1 core @ >3.5 Ghz (C/C++)
J. Revaud, P. Weinzaepfel, Z. Harchaoui and C. Schmid: EpicFlow: Edge-Preserving
Interpolation of Correspondences for Optical
Flow . CVPR 2015 - IEEE Conference on
Computer Vision \& Pattern Recognition 2015.
231
SegFlow(d0=11)
28.97 %
22.64 %
27.91 %
100.00 %
4.5 s
1 core @ 3.5 Ghz (C/C++)
J. Chen, Z. Cai, J. Lai and X. Xie: Efficient Segmentation-based PatchMatch for
Large displacement Optical
Flow Estimation . IEEE TCSVT 2018.
232
DeepFlow
code
27.96 %
31.06 %
28.48 %
100.00 %
17 s
1 core @ >3.5 Ghz (Python + C/C++)
P. Weinzaepfel, J. Revaud, Z. Harchaoui and C. Schmid: DeepFlow: Large displacement optical
flow with deep matching . IEEE Intenational Conference on
Computer Vision (ICCV) 2013.
233
CPNFlow
31.05 %
27.16 %
30.40 %
100.00 %
0.1 s
GPU @ 1.5 Ghz (Python)
Y. Yang and S. Soatto: Conditional prior networks for optical flow . Proceedings of the European Conference on Computer
Vision (ECCV) 2018.
234
IIOF-NLDP
30.23 %
32.44 %
30.60 %
100.00 %
350 s
4 cores @ 3.5 Ghz (Matlab + C/C++)
D. Trinh, W. Blondel and C. Daul: A General Form of Illumination-
Invariant Descriptors in Variational Optical Flow
Estimation . IEEE Int. Conf. on Image
Processing (ICIP) 2017.
235
DMF_ROB
code
30.74 %
30.07 %
30.63 %
100.00 %
150 s
1 core @ 2.5 Ghz (C/C++)
P. Weinzaepfel, J. Revaud, Z. Harchaoui and C. Schmid: DeepFlow: Large displacement optical flow with deep matching . ICCV - IEEE International Conference on Computer Vision 2013.
236
SPyNet
code
33.36 %
43.62 %
35.07 %
100.00 %
0.16 s
1 core @ 2.5 Ghz (C/C++)
A. Ranjan and M. Black: Optical Flow Estimation using a
Spatial Pyramid Network . Proceedings of the IEEE Conference
on Computer Vision and Pattern Recognition 2017.
237
SGM+C+NL
code
34.24 %
42.46 %
35.61 %
93.83 %
4.5 min
1 core @ 2.5 Ghz (C/C++)
H. Hirschmüller: Stereo Processing by Semiglobal Matching and Mutual Information . PAMI 2008. D. Sun, S. Roth and M. Black: A Quantitative Analysis of Current Practices in Optical Flow Estimation and the Principles Behind Them . IJCV 2013.
238
DWBSF
40.74 %
31.16 %
39.14 %
100.00 %
7 min
4 cores @ 3.5 Ghz (C/C++)
C. Richardt, H. Kim, L. Valgaerts and C. Theobalt: Dense Wide-Baseline Scene Flow
From Two Handheld Video Cameras . 3DV 2016.
239
SGM+LDOF
code
40.81 %
31.92 %
39.33 %
95.89 %
86 s
1 core @ 2.5 Ghz (C/C++)
H. Hirschmüller: Stereo Processing by Semiglobal Matching and Mutual Information . PAMI 2008. T. Brox and J. Malik: Large Displacement Optical Flow: Descriptor Matching in Variational Motion Estimation . PAMI 2011.
240
3DG-DVO
35.78 %
63.65 %
40.42 %
100.00 %
0.04 s
GPU @ 1.5 Ghz (Python)
241
HS
code
39.90 %
51.39 %
41.81 %
100.00 %
2.6 min
1 core @ 3.0 Ghz (Matlab)
D. Sun, S. Roth and M. Black: A Quantitative Analysis of Current
Practices in Optical Flow Estimation
and The Principles Behind Them . 2014.
242
GCSF
code
47.38 %
41.50 %
46.40 %
100.00 %
2.4 s
1 core @ 2.5 Ghz (C/C++)
J. Cech, J. Sanchez-Riera and R. Horaud: Scene Flow Estimation by growing Correspondence Seeds . CVPR 2011.
243
DB-TV-L1
code
47.52 %
48.27 %
47.64 %
100.00 %
16 s
1 core @ 2.5 Ghz (Matlab)
C. Zach, T. Pock and H. Bischof: A Duality Based Approach for Realtime TV-
L1 Optical Flow . DAGM 2007.
244
VSF
code
50.06 %
45.40 %
49.28 %
100.00 %
125 min
1 core @ 2.5 Ghz (C/C++)
F. Huguet and F. Devernay: A Variational Method for Scene Flow Estimation from Stereo Sequences . ICCV 2007.
245
HAOF
code
49.89 %
50.74 %
50.04 %
100.00 %
16.2 s
1 core @ 2.5 Ghz (C/C++)
T. Brox, A. Bruhn, N. Papenberg and J. Weickert: High accuracy optical flow estimation
based on a theory for warping . ECCV 2004.
246
TVL1_ROB
code
51.15 %
51.12 %
51.14 %
100.00 %
3 s
4 cores @ 2.5 Ghz (C/C++)
J. Sánchez Pérez, E. Meinhardt-Llopis and G. Facciolo: TV-L1 Optical Flow Estimation . Image Processing On Line 2013.
247
PolyExpand
52.00 %
58.56 %
53.09 %
100.00 %
1 s
1 core @ 2.5 Ghz (C/C++)
G. Farneback: Two-Frame Motion Estimation Based on Polynomial Expansion . SCIA 2003.
248
H+S_ROB
code
68.22 %
76.49 %
69.60 %
100.00 %
8 s
4 cores @ 2.5 Ghz (C/C++)
E. Meinhardt-Llopis, J. Sánchez Pérez and D. Kondermann: Horn-Schunck Optical Flow with a Multi-Scale Strategy . Image Processing On Line 2013.
249
Stereo-RSSF
code
70.68 %
73.60 %
71.17 %
9.26 %
2.5 s
8 core @ 2.5 Ghz (Matlab)
E. Salehi, A. Aghagolzadeh and R. Hosseini: Stereo-RSSF: stereo robust sparse scene-flow estimation . The Visual Computer 2023.
250
Pyramid-LK
code
71.84 %
76.82 %
72.67 %
100.00 %
1.5 min
1 core @ 2.5 Ghz (Matlab)
J. Bouguet: Pyramidal implementation of the Lucas
Kanade feature tracker . Intel 2000.
251
MEDIAN
87.37 %
92.80 %
88.27 %
99.86 %
0.01 s
1 core @ 2.5 Ghz (C/C++)
252
AVERAGE
88.47 %
92.08 %
89.07 %
99.86 %
0.01 s
1 core @ 2.5 Ghz (C/C++)